Fiber/coax networks (FCNs) that combine the low cost and large bandwidth of coaxial cable with the flexibility of lightwave feeder technology have been shown to be a very promising system architecture for current and near-future broadband local access applications. Optical fiber is used to span large distances from the central office (CO) to the remote fiber node (FN) where the coaxial cable distributes information to the end users (EUs). FCNs offer an immediate low-cost path to any presently defined broadband or narrowband, broadcast or switched, analog or digital services and can be upgraded to provide increased bandwidth for interactive multimedia services or other future services. For cable TV companies, the advantages of this architecture have already been demonstrated where the migration from pure coax to fiber/coax networks has resulted in improved signal quality, higher reliability and greatly increased bandwidth (approaching 1 GHz) to the homes. For local exchange carriers (LECs), FCNs provide sufficient bandwidth for video services for less cost than alternative subscriber loop systems. With its low cost, large bandwidth, and high penetration (95% of U.S. homes have been passed by existing cable TV coax networks), FCNs are presently the most popular broadband access infrastructure for current and near-future information services.
To realize the full potential of FCNs, more technical innovation is required. One of the most important challenges is to cost-effectively provide broadband two-way services over a system that has been designed primarily for distributed broadcast television services. Proposed FCNs support limited two-way services by defining the upstream traffic in the traditional upstream frequency band of 5-40 MHz as shown by band Upstream 1 in FIG. 1. However, this small upstream bandwidth limits services that can be provided. Further, due to heavy in-the-air traffic (i.e., amateur radio) in that frequency range, ingress noise in the coaxial cable seriously degrades channel performance.
To overcome this ingress noise and create more upstream bandwidth, one approach is to use a high-frequency split plan, where the upstream traffic is located in one band at frequencies greater than the downstream band, as shown by Upstream 2 in FIG. 1. This approach does not affect existing downstream services and has low ingress noise in the broadband return path. However, both this approach and the traditional approach have the limitation that the total bandwidth has to be pre-divided into downstream and upstream bands, with diplexers (or triplexers) and separate upstream amplifiers installed in all amplifiers to provide non-overlapping bi-directional paths. Downstream and upstream bandwidth allocations are then restricted to those defined during construction and cannot be changed without fully renovating the coax network. This fixed pre-provision frequency plan limits the network's capability to support wide varieties of future broadband two-way symmetric and asymmetric services.